tychonic 0.2.3rc4__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- tychonic-0.2.3rc4/CHANGELOG.md +153 -0
- tychonic-0.2.3rc4/Cargo.lock +1275 -0
- tychonic-0.2.3rc4/Cargo.toml +42 -0
- tychonic-0.2.3rc4/LICENSE-APACHE +201 -0
- tychonic-0.2.3rc4/LICENSE-MIT +21 -0
- tychonic-0.2.3rc4/PKG-INFO +71 -0
- tychonic-0.2.3rc4/README.md +259 -0
- tychonic-0.2.3rc4/benches/kernels.rs +236 -0
- tychonic-0.2.3rc4/benches/support/mod.rs +498 -0
- tychonic-0.2.3rc4/clippy.toml +12 -0
- tychonic-0.2.3rc4/docs/acquisition.md +160 -0
- tychonic-0.2.3rc4/docs/advanced_api.md +315 -0
- tychonic-0.2.3rc4/docs/batch-covariance-requirements.md +41 -0
- tychonic-0.2.3rc4/docs/constrained_range_rate.md +133 -0
- tychonic-0.2.3rc4/docs/covariance.md +78 -0
- tychonic-0.2.3rc4/docs/dart-compatibility-0.2.3.md +53 -0
- tychonic-0.2.3rc4/docs/differential_doppler.md +277 -0
- tychonic-0.2.3rc4/docs/migration_0_9_2.md +77 -0
- tychonic-0.2.3rc4/docs/migration_0_9_7.md +80 -0
- tychonic-0.2.3rc4/docs/phase_difference.md +173 -0
- tychonic-0.2.3rc4/docs/publishing.md +122 -0
- tychonic-0.2.3rc4/docs/python-api.md +395 -0
- tychonic-0.2.3rc4/docs/python-low-level.md +272 -0
- tychonic-0.2.3rc4/docs/python.md +94 -0
- tychonic-0.2.3rc4/docs/qualification_0_9_2.json +841 -0
- tychonic-0.2.3rc4/docs/qualification_0_9_2.md +120 -0
- tychonic-0.2.3rc4/docs/release_0_9_3.md +130 -0
- tychonic-0.2.3rc4/docs/safety.md +138 -0
- tychonic-0.2.3rc4/docs/simulation.md +24 -0
- tychonic-0.2.3rc4/examples/acquisition.rs +39 -0
- tychonic-0.2.3rc4/examples/batch_fit.rs +129 -0
- tychonic-0.2.3rc4/examples/differential_doppler.rs +26 -0
- tychonic-0.2.3rc4/examples/passive_fusion.rs +127 -0
- tychonic-0.2.3rc4/examples/rf_offsets.rs +85 -0
- tychonic-0.2.3rc4/examples/support/acquisition.rs +265 -0
- tychonic-0.2.3rc4/pyproject.toml +62 -0
- tychonic-0.2.3rc4/python/Cargo.lock +1267 -0
- tychonic-0.2.3rc4/python/Cargo.toml +27 -0
- tychonic-0.2.3rc4/python/README.md +41 -0
- tychonic-0.2.3rc4/python/examples/acquire.py +83 -0
- tychonic-0.2.3rc4/python/examples/contact_reference.rs +21 -0
- tychonic-0.2.3rc4/python/examples/fit_and_fuse.py +53 -0
- tychonic-0.2.3rc4/python/examples/handoff_reference.rs +169 -0
- tychonic-0.2.3rc4/python/examples/reference.rs +286 -0
- tychonic-0.2.3rc4/python/examples/tca_correction.py +44 -0
- tychonic-0.2.3rc4/python/src/common.rs +192 -0
- tychonic-0.2.3rc4/python/src/config.rs +78 -0
- tychonic-0.2.3rc4/python/src/lib.rs +50 -0
- tychonic-0.2.3rc4/python/src/measurement.rs +273 -0
- tychonic-0.2.3rc4/python/src/mpc.rs +310 -0
- tychonic-0.2.3rc4/python/src/orbit.rs +437 -0
- tychonic-0.2.3rc4/python/src/profiles.rs +269 -0
- tychonic-0.2.3rc4/python/src/solve.rs +292 -0
- tychonic-0.2.3rc4/python/src/tca.rs +215 -0
- tychonic-0.2.3rc4/python/tests/conftest.py +53 -0
- tychonic-0.2.3rc4/python/tests/data/contact_reference.json +1 -0
- tychonic-0.2.3rc4/python/tests/data/handoff_reference.json +1 -0
- tychonic-0.2.3rc4/python/tests/data/ksat_a2.tdm +81 -0
- tychonic-0.2.3rc4/python/tests/data/reference.json +1688 -0
- tychonic-0.2.3rc4/python/tests/data/two_contacts.json +1 -0
- tychonic-0.2.3rc4/python/tests/test_acquisition.py +105 -0
- tychonic-0.2.3rc4/python/tests/test_estimation.py +219 -0
- tychonic-0.2.3rc4/python/tests/test_inputs.py +99 -0
- tychonic-0.2.3rc4/python/tests/test_tdm.py +231 -0
- tychonic-0.2.3rc4/python/tests/test_workflows.py +584 -0
- tychonic-0.2.3rc4/python/tychonic/__init__.py +9 -0
- tychonic-0.2.3rc4/python/tychonic/_catalogue.py +295 -0
- tychonic-0.2.3rc4/python/tychonic/_exports.py +94 -0
- tychonic-0.2.3rc4/python/tychonic/_native.pyi +200 -0
- tychonic-0.2.3rc4/python/tychonic/_profiles.py +308 -0
- tychonic-0.2.3rc4/python/tychonic/_search.py +92 -0
- tychonic-0.2.3rc4/python/tychonic/_tdm.py +209 -0
- tychonic-0.2.3rc4/python/tychonic/_tdm_export.py +230 -0
- tychonic-0.2.3rc4/python/tychonic/_utils.py +61 -0
- tychonic-0.2.3rc4/python/tychonic/acquisition.py +221 -0
- tychonic-0.2.3rc4/python/tychonic/api.py +1305 -0
- tychonic-0.2.3rc4/python/tychonic/config.py +126 -0
- tychonic-0.2.3rc4/python/tychonic/estimation.py +298 -0
- tychonic-0.2.3rc4/python/tychonic/low_level.py +85 -0
- tychonic-0.2.3rc4/python/tychonic/measurements.py +246 -0
- tychonic-0.2.3rc4/python/tychonic/py.typed +0 -0
- tychonic-0.2.3rc4/references/README.md +12 -0
- tychonic-0.2.3rc4/src/detection/belief.rs +180 -0
- tychonic-0.2.3rc4/src/detection/geometry.rs +225 -0
- tychonic-0.2.3rc4/src/detection/mod.rs +80 -0
- tychonic-0.2.3rc4/src/detection/mpc.rs +400 -0
- tychonic-0.2.3rc4/src/detection/planner.rs +437 -0
- tychonic-0.2.3rc4/src/detection/types.rs +235 -0
- tychonic-0.2.3rc4/src/estimate/ekf.rs +784 -0
- tychonic-0.2.3rc4/src/estimate/history.rs +61 -0
- tychonic-0.2.3rc4/src/estimate/mod.rs +10 -0
- tychonic-0.2.3rc4/src/estimate/passive_rf.rs +578 -0
- tychonic-0.2.3rc4/src/estimate/tca_ukf.rs +656 -0
- tychonic-0.2.3rc4/src/kernels/mod.rs +143 -0
- tychonic-0.2.3rc4/src/kernels/observation_models.rs +738 -0
- tychonic-0.2.3rc4/src/kernels/phase.rs +340 -0
- tychonic-0.2.3rc4/src/kernels/rotation.rs +629 -0
- tychonic-0.2.3rc4/src/kernels/stm.rs +682 -0
- tychonic-0.2.3rc4/src/lib.rs +37 -0
- tychonic-0.2.3rc4/src/measurements/rf.rs +158 -0
- tychonic-0.2.3rc4/src/measurements.rs +800 -0
- tychonic-0.2.3rc4/src/optim/constrained/initialization.rs +99 -0
- tychonic-0.2.3rc4/src/optim/constrained/mod.rs +419 -0
- tychonic-0.2.3rc4/src/optim/constrained/selection.rs +130 -0
- tychonic-0.2.3rc4/src/optim/constrained/solver.rs +380 -0
- tychonic-0.2.3rc4/src/optim/covariance.rs +244 -0
- tychonic-0.2.3rc4/src/optim/input.rs +28 -0
- tychonic-0.2.3rc4/src/optim/lm.rs +431 -0
- tychonic-0.2.3rc4/src/optim/mod.rs +348 -0
- tychonic-0.2.3rc4/src/optim/passive_rf.rs +730 -0
- tychonic-0.2.3rc4/src/optim/tle_correction.rs +143 -0
- tychonic-0.2.3rc4/src/optim/varpro.rs +474 -0
- tychonic-0.2.3rc4/src/simulation.rs +365 -0
- tychonic-0.2.3rc4/tests/acquisition.rs +1068 -0
- tychonic-0.2.3rc4/tests/batch.rs +618 -0
- tychonic-0.2.3rc4/tests/bootstrap.rs +757 -0
- tychonic-0.2.3rc4/tests/constrained_fit.rs +303 -0
- tychonic-0.2.3rc4/tests/covariance.rs +404 -0
- tychonic-0.2.3rc4/tests/differential.rs +680 -0
- tychonic-0.2.3rc4/tests/ekf.rs +574 -0
- tychonic-0.2.3rc4/tests/geometry.rs +175 -0
- tychonic-0.2.3rc4/tests/kernel_stress.rs +225 -0
- tychonic-0.2.3rc4/tests/measurements.rs +231 -0
- tychonic-0.2.3rc4/tests/passive_api.rs +590 -0
- tychonic-0.2.3rc4/tests/phase.rs +510 -0
- tychonic-0.2.3rc4/tests/prepared.rs +411 -0
- tychonic-0.2.3rc4/tests/propagation.rs +408 -0
- tychonic-0.2.3rc4/tests/simulation.rs +115 -0
- tychonic-0.2.3rc4/tests/support/constrained.rs +131 -0
- tychonic-0.2.3rc4/tests/support/mod.rs +116 -0
- tychonic-0.2.3rc4/tests/tca_ukf.rs +525 -0
- tychonic-0.2.3rc4/tests/variable_baselines.rs +681 -0
|
@@ -0,0 +1,153 @@
|
|
|
1
|
+
# Changelog
|
|
2
|
+
|
|
3
|
+
## Python 0.2.3rc4
|
|
4
|
+
|
|
5
|
+
- `ContactBiasFit` accepts independent gradient, step and cost tolerances. Each omitted tolerance uses its native LM default. The contact solver factors the projected Jacobian to return an exactly symmetric residual-scaled Schur covariance for EKF initialization.
|
|
6
|
+
- Cartesian TLE and OMM exports accept a NORAD catalog ID and international designator. Fitted OMM epochs use UTC with microsecond precision and omit the trailing `Z` for interoperability with `sgp4.omm.initialize`.
|
|
7
|
+
- OEM exports include the stored GCRF orbital covariance at the estimate epoch when it lies within the sampled interval.
|
|
8
|
+
- Add a batch SGP4 fit for TLE epoch offset, offset rate, and Doppler bias from joint Doppler and pointing observations. Python callers can use `fit_tle_correction()` with `TrackingData` (also available as `JobData`) and export the correction at a handoff time. [Python API](docs/python-api.md#tlecorrection)
|
|
9
|
+
|
|
10
|
+
## Python 0.2.1
|
|
11
|
+
|
|
12
|
+
- Export one-way range rate, KSAT A2 pointing angles, and CCSDS differential Doppler as separate KVN strings from `JobData.to_ccsds_tdm()`. The directory writer accepts the same product selection and defaults to range rate. [TDM conventions](docs/python-api.md#tracking-data-messages)
|
|
13
|
+
|
|
14
|
+
## 0.10.1 (release candidate)
|
|
15
|
+
|
|
16
|
+
- Constrained fitting is included in `tychonic`. Remove the `tychonic-constrained` dependency and change Rust imports from `tychonic_constrained::` to `tychonic::optim::constrained::`. The fitting function and its arguments are unchanged.
|
|
17
|
+
- Calibrated LM accepts an explicit independent Gaussian prior. Fitted state and full covariance can be transported to a publication epoch with the native variational model.
|
|
18
|
+
- Add fixed-drag Cartesian EKFs with additive Doppler bias in Hz and TCA UKFs with per-reading carrier and correlated Doppler/pointing noise. Existing drag and range-rate filter constructors remain available.
|
|
19
|
+
- Add the independently versioned Python 0.2.0 mixed Rust/Python distribution, including all four batch profiles, per-spacecraft checkpoint/restore, epoch-linked RTS smoothing, MPC and CCSDS exports. See the [Python guide](docs/python.md).
|
|
20
|
+
|
|
21
|
+
## 0.10.0 (unreleased)
|
|
22
|
+
|
|
23
|
+
- Add a prepared seven-variable range-rate NLP with projected pass biases, frozen closest-approach sampling and hard circular field-of-view bounds.
|
|
24
|
+
- Add the opt-in `tychonic-constrained` Rust solver crate using ripopt 0.8.2, and native IPOPT qualification. See the [fitting guide](docs/constrained_range_rate.md).
|
|
25
|
+
|
|
26
|
+
## 0.9.7 (2026-11-11)
|
|
27
|
+
|
|
28
|
+
- **Breaking:** move surveyed baselines out of `Interferometer`; construction now
|
|
29
|
+
takes only the nominal carrier frequency. Phase-difference and differential
|
|
30
|
+
Doppler measurements require explicit ENU baselines, with one vector per
|
|
31
|
+
measurement in batch inputs. Fixed baselines use repeated vectors.
|
|
32
|
+
- Carry per-measurement baselines through joint observations, batch sensitivity
|
|
33
|
+
mappings, RF fitting, and chronologically sorted streaming updates.
|
|
34
|
+
- Keep shared constant ENU calibration corrections, applied to every supplied
|
|
35
|
+
baseline vector, and independent corrections for separate differential streams.
|
|
36
|
+
- Update examples, tools, and [migration guidance](docs/migration_0_9_7.md).
|
|
37
|
+
- Verify varying baselines, duplicate and unsorted epochs, analytic derivatives,
|
|
38
|
+
zero/reversed baselines, invalid inputs, and retained fixed-baseline behavior.
|
|
39
|
+
|
|
40
|
+
## 0.9.6
|
|
41
|
+
|
|
42
|
+
- Consolidate the 0.9.4 absolute calibrated covariance APIs and 0.9.5 behavioral
|
|
43
|
+
antenna/receiver kernels in the canonical upstream checkout.
|
|
44
|
+
- Align the development-tools dependency lockfile with the library version.
|
|
45
|
+
- Anchor source-package entries to the repository root, excluding local
|
|
46
|
+
Python-environment files while retaining the citation index.
|
|
47
|
+
- Update release metadata and document local release preparation.
|
|
48
|
+
|
|
49
|
+
## 0.9.5
|
|
50
|
+
|
|
51
|
+
- Add reusable caller-clocked behavioral mount and receiver simulation kernels.
|
|
52
|
+
- Add Gaussian beam loss, hard reception cutoff, acquisition dwell, lock
|
|
53
|
+
hysteresis, seeded Doppler noise, and absent unlocked observations.
|
|
54
|
+
- Add actual-epoch orbital frequency/pointing samples and cache-safe TLE epoch
|
|
55
|
+
hypotheses. Keep transport and spacecraft-truth disclosure caller-owned.
|
|
56
|
+
- Verify cutoff boundaries, actuator limits, seeded replay, invalid inputs,
|
|
57
|
+
frequency against independent range differences, and retained numerical APIs.
|
|
58
|
+
|
|
59
|
+
## 0.9.4
|
|
60
|
+
|
|
61
|
+
- Add covariance-enabled calibrated batch LM and facade APIs without changing
|
|
62
|
+
existing fit functions or `FitError` variants. Return physical parameter
|
|
63
|
+
covariance, numerical rank, and its statistical interpretation.
|
|
64
|
+
- Compute absolute whitened-information covariance from the final retained LM
|
|
65
|
+
evaluation using column-normalized pivoted QR and compact R SVD. Exclude
|
|
66
|
+
residual scaling and solver damping; retain Varpro Schur covariance behavior.
|
|
67
|
+
- Return structured rank and numerical errors instead of regularization or
|
|
68
|
+
pseudo-priors. Keep covariance storage linear in observation count.
|
|
69
|
+
- Verify analytic correlated-noise references, units and optimizer-scale
|
|
70
|
+
invariance, six/seven-parameter fits, singular geometry, arithmetic overflow,
|
|
71
|
+
and noisy mixed-observable recovery. Add a reproducible downstream fixture.
|
|
72
|
+
|
|
73
|
+
## 0.9.3
|
|
74
|
+
|
|
75
|
+
Release contents include all additions since 0.9.2. See the
|
|
76
|
+
[feature outline](docs/release_0_9_3.md) for interfaces, integration, and validation.
|
|
77
|
+
|
|
78
|
+
- Add the public `detection` module and `BcMpc` controller for Bayesian
|
|
79
|
+
Continuation MPC satellite acquisition from explicit weighted TLE/OMM hypotheses.
|
|
80
|
+
- Build a private feasible ATUS continuation, then optimize a moving candidate
|
|
81
|
+
prefix against acquisition probability over the entire remaining search.
|
|
82
|
+
Preserve the continuation boundary and retain the incumbent plan when beam
|
|
83
|
+
search cannot improve its score.
|
|
84
|
+
- Add particle-belief normalization, log-space Bayesian miss updates, effective
|
|
85
|
+
sample size, and caller-seeded systematic resampling. Cache geometry for the
|
|
86
|
+
original ephemeris pool and select active hypotheses by ancestor indices.
|
|
87
|
+
- Add ENU pointing geometry, along-track candidates, elevation-mask checks,
|
|
88
|
+
rest-to-rest antenna slew constraints, and partial slews for initialization.
|
|
89
|
+
- Add timed actions, acquisition/miss observations, explicit ready/pending/
|
|
90
|
+
acquired/exhausted states, and a caller-supplied `DetectionModel`. Planning
|
|
91
|
+
failures, unsupported misses, and invalid resampling draws preserve controller
|
|
92
|
+
state; execution failures leave the action pending rather than reporting a miss.
|
|
93
|
+
- Add checked batch `kernels::propagate_sgp4_gcrf` for any Satkit `SGP4Source`,
|
|
94
|
+
including TLE and OMM, preserving epoch order and checking every propagation
|
|
95
|
+
status, output shape, frame conversion, and state. Reuse it in the existing
|
|
96
|
+
TLE initializer and combine it with prepared station geometry for acquisition.
|
|
97
|
+
- Add a shared simulated plant and runnable `acquisition` example with an
|
|
98
|
+
independent true orbit, erroneous planning orbit, Gaussian receiver model,
|
|
99
|
+
separate seeded receiver/resampling streams, and per-cycle diagnostics.
|
|
100
|
+
The default fixture discovers the satellite in 12 simulated seconds.
|
|
101
|
+
- Add acquisition regressions for kernel/reference agreement, constrained
|
|
102
|
+
planning, belief/resampling behavior, state transitions and error atomicity,
|
|
103
|
+
TLE/OMM geometry, reproducible simulated discovery, budget exhaustion, and
|
|
104
|
+
failed-execution retry behavior. Retain the forced-feedback controller test.
|
|
105
|
+
- Add the acquisition API/defaults guide, README integration example, simulated
|
|
106
|
+
discovery instructions, and the satellite-search paper reference.
|
|
107
|
+
|
|
108
|
+
## 0.9.2
|
|
109
|
+
|
|
110
|
+
- Replace reference counting and interior mutability in Tychonic-owned code
|
|
111
|
+
with owned or borrowed geometry, observation data and problems, plus reusable
|
|
112
|
+
mutable evaluation buffers. Kernels execute sequentially on the caller's thread.
|
|
113
|
+
- Add a local fallible LM driver for ordinary, RF and variable-projection fits,
|
|
114
|
+
preserving damping and convergence equations while evaluating residuals and
|
|
115
|
+
Jacobians together and rejecting negative-drag trials before propagation.
|
|
116
|
+
- Validate sensitivity arrays, RF data, priors, epochs and integration controls
|
|
117
|
+
before use. Return numerical and solver errors without panic extraction.
|
|
118
|
+
- Store filter history with an always-present current record; restore small
|
|
119
|
+
checkpoints on failed observations and materialize RTS records explicitly.
|
|
120
|
+
- Retain owned covariance factors and share one factor across independent joint
|
|
121
|
+
rows. Remove whole-estimator and whole-evaluation cloning.
|
|
122
|
+
- Forbid unsafe library code and enforce ownership/error policies with Clippy,
|
|
123
|
+
a source check and CI. Pin and document the Numeris/Satkit trust boundary and
|
|
124
|
+
narrowly scoped external settings/TLE copies.
|
|
125
|
+
- Intentionally break ownership, validated-input and history APIs; update
|
|
126
|
+
examples, tests and tools. See [migration](docs/migration_0_9_2.md),
|
|
127
|
+
[safety contracts](docs/safety.md) and [qualification](docs/qualification_0_9_2.md).
|
|
128
|
+
|
|
129
|
+
- Add a two-state TCA UKF for TLE epoch correction and a constant pass
|
|
130
|
+
range-rate bias, using combined range-rate and soft azimuth/elevation readings.
|
|
131
|
+
- Default TCA pointing noise to independent 2.5-degree standard deviations;
|
|
132
|
+
retain observation-time station geometry and invalidate shifted SGP4 caches.
|
|
133
|
+
|
|
134
|
+
## 0.9.0
|
|
135
|
+
|
|
136
|
+
- Add a calibrated facade for standalone Doppler Hz, continuous phase difference,
|
|
137
|
+
exact differential Doppler, and azimuth/elevation.
|
|
138
|
+
- Fuse any selection with full per-epoch covariance in batch fits and EKF
|
|
139
|
+
updates, retaining shared ordinary/differential Doppler noise.
|
|
140
|
+
- Add fit convenience functions, named orbit filter aliases, and an RF builder
|
|
141
|
+
with named priors for existing offset, drift and baseline-calibration models.
|
|
142
|
+
- Preserve analytic Jacobians, prepared geometry, propagation, variable
|
|
143
|
+
projection, augmented RF estimators and RTS smoothing.
|
|
144
|
+
- Preserve existing public functions; extend `MeasurementKind` with `Joint`.
|
|
145
|
+
Downstream exhaustive matches must handle the new variant.
|
|
146
|
+
- Move research/profiling executables and plotting scripts to the development
|
|
147
|
+
tools crate. Replace the default executable with the `batch_fit` example.
|
|
148
|
+
- Keep generated experiments and reference PDFs locally, outside Git tracking
|
|
149
|
+
and release packages. Add MIT/Apache-2.0 licensing and release metadata.
|
|
150
|
+
- Clarify that fitting requires an initial orbit guess; this release does not
|
|
151
|
+
add orbit initialization without a guess or receiver signal extraction.
|
|
152
|
+
|
|
153
|
+
Earlier experimental implementations are not assigned release versions here.
|